The arguments put forward by Peters et al. are certainly well thought out and presented in a logical fashion. Although there is much to agree upon, there are also points that need further discussion.
In their introduction, Peters et al. refer to simple versus complex carotid intima-media thickness (CIMT) imaging and processing protocols and state “that these extensive protocols are required to obtain the highest precision to observe a treatment effect and to fully cover the degree of atherosclerotic burden.”
A response to this comment is rapidly formulated. Have not other intima-media thickness (IMT) protocols, much simpler than the Measuring Effects on Intima-Media Thickness: An Evaluation of Rosuvastatin (METEOR) protocol, shown positive associations between interventions and change in IMT? The latter meta-analysis did not adjust for ultrasound protocols. As to the “highest precision,” data published by METEOR have shown that restriction of analyses to only one segment and to only one projection have similar variability for determining change in IMT as those derived from multiple angles and multiple projections.
With this as a background, it is of interest to review the points made by Peters et al.
Completeness of Carotid Intima-Media Thickness Information
The statement that “near wall measurements and measurements from the carotid bifurcation and internal carotid artery are commonly thought to be difficult to obtain, and therefore give rise to missing data” is of interest.
Although it is true that certain investigators have promulgated the belief that near-wall CIMT measurements are unreliable and difficult to obtain, this was never an issue put forward in the original editorial comment of their work. Near-wall CIMT measurements have been an integral part of many population-based epidemiologic studies, such as the Cardiovascular Health Study, the Coronary Artery Risk in Young Adults study, and the Framingham Heart Study.
The major point of this discussion remains whether there is a net benefit to making the CIMT imaging protocol a complex one.
Precision and Magnitude of Estimated Rates of Change
The investigators’ arguments about decreasing heterogeneity of maximum CIMT by adopting multiple-angle sampling seem unsupported. The data presented in their Table 5 do not show a significant difference in precision between maximum IMT estimated from one compared with five angles. If there were a significant decrease in precision for the one-angle protocol, the data in the table should reflect this.
Peters et al. point out that estimated progression rates are larger in the bifurcation than in the common carotid artery. This is well known and likely due to different associations between risk factors and IMT in these segments. Would combining measurements of IMT rates of change from different segments (“lumping”) preserve these associations? Combining segmental information to create one global estimate might dilute potential associations between IMT rate of change and specific risk factors.
The asymmetric nature of atherosclerotic plaque formation is well known in clinical carotid artery imaging. It would therefore seem logical that the sonographer performing a clinical IMT evaluation be directed to acquire the best projection on which maximum IMT (in essence plaque) is shown. This approach seems to work well in epidemiologic studies and for protocols in which plaque area is measured as part of cross-sectional and longitudinal studies.
Duration of the Acquisition Protocol
A sonographer would respond to this argument by pointing out that clinical ultrasound examinations take approximately 30 min and are not complicated by the placement of a device that forces the sonographer to image in sometimes compromising positions. The latter possibility is a serious one to take into consideration, because ergonomics play a very important role in causing musculoskeletal injuries in sonographers.
A report that shares two primary authors with the article by Peters et al. reviews the reproducibility of IMT measurements using a three-segment and four-angle protocol very similar to the one adopted by METEOR. The CIMT measurement software was automated, making it more efficient than manually traced IMT measurements, so the image processing time should at least be equal if not less than the process used in METEOR. A one-segment, one-angle IMT study was estimated to take 15 min for acquisition, compared with 45 min for a three-segment, four-angle protocol. A five-angle protocol would therefore likely take approximately 56 min to acquire. IMT measurement time was estimated at 15 min for a one-segment protocol but between 90 and 120 min for a three-segment, four-angle protocol. Estimated processing time for a five-angle protocol would therefore be 110 to 150 min. Adding these numbers indicates that the total time taken to acquire and process images would be 30 min for a one-segment protocol and 156 to 206 min for a five-segment, five-angle protocol.
The comparison is between 0.5 hours for a simple CIMT protocol and 3 hours for the three-segment, five-angle protocol. This should be weighed against the observation that in METEOR, the one-segment, one-angle protocol gave similar results to the three-segment, five-angle protocol (Table 5 in Peters et al. ). Despite the much larger investment in time and resources, the hypothesized improvement in precision is not significant.
Precision and Magnitude of Estimated Rates of Change
The investigators’ arguments about decreasing heterogeneity of maximum CIMT by adopting multiple-angle sampling seem unsupported. The data presented in their Table 5 do not show a significant difference in precision between maximum IMT estimated from one compared with five angles. If there were a significant decrease in precision for the one-angle protocol, the data in the table should reflect this.
Peters et al. point out that estimated progression rates are larger in the bifurcation than in the common carotid artery. This is well known and likely due to different associations between risk factors and IMT in these segments. Would combining measurements of IMT rates of change from different segments (“lumping”) preserve these associations? Combining segmental information to create one global estimate might dilute potential associations between IMT rate of change and specific risk factors.
The asymmetric nature of atherosclerotic plaque formation is well known in clinical carotid artery imaging. It would therefore seem logical that the sonographer performing a clinical IMT evaluation be directed to acquire the best projection on which maximum IMT (in essence plaque) is shown. This approach seems to work well in epidemiologic studies and for protocols in which plaque area is measured as part of cross-sectional and longitudinal studies.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
